Ravi Jagadeesan, Mentor: Akhil Mathew MIT PRIMES May 18, 2013
Compact Riemann surfaces Definition A Riemann surface is a one-dimensional complex manifold. 1 1 http://upload.wikimedia.org/wikipedia/commons/f/f0/triple_ torus_illustration.png.
Algebraic curves Loosely speaking, an algebraic curve is a one-dimensional object that is the set of common zeros of a finite set of multivariate polynomials. Theorem (Riemann Existence) y 2 = x 3 x Every compact Riemann surface has an algebraic structure. An algebraic curve is defined over Q if the equations can be taken to have coefficients in Q.
Belyi functions Definition A Belyi function is a morphism f : X P 1 C, X a compact Riemann surface (smooth, projective, irreducible curve over C) f unbranched outside {0, 1, }. Theorem (Belyi) An algebraic curve over C admits a Belyi function if and only if it is defined over Q.
Dessins d enfants Definition A dessin d enfant is a connected, bipartite graph G embedded as a map into a topological compact oriented surface. 2 M. M. Wood, Belyi-Extending Maps and the Galois Action on Dessins d Enfants, Publ. RIMS, Kyoto Univ., 42: 721737, 2006. 2
Correspondence between Belyi functions and dessins Theorem (Grothendieck) There is a natural one-to-one correspondence between isomorphism classes of Belyi functions and isomorphism classes of dessins d enfants. Associate f : X P 1 C to the graph f 1 ([0, 1]). Bipartite with parts V 0 = f 1 ({0}) and V 1 = f 1 ({1})
The full dictionary of objects The following objects define equivalent data: isomorphism classes of dessin d enfants with n edges isomorphism classes of Belyi functions of degree n conjugacy classes of transitive representations x, y S n conjugacy classes of subgroups of x, y of index n
Action of Gal(Q/Q) There is a natural action of Gal(Q/Q) on the category of algebraic curves over Q. Apply an automorphism of Q to all the coefficients of the defining polynomials Equivalently, base-change by an automorphism of Spec Q By Belyi s Theorem, this gives an action of Gal(Q/Q) on the category of Belyi functions. The action is faithful. Hence, Gal(Q/Q) acts faithfully on the set of isomorphism classes of dessins as well. So, a number-theoretic object acts on a purely combinatorial object.
Application to inverse Galois theory Question (Inverse Galois Problem) What are all finite quotients of Gal(Q/Q)? What is Gal(Q/Q)? One can study Gal(Q/Q) by its (faithful) action on the category of dessins. Question (Grothendieck) How does Gal(Q/Q) act on the category of Belyi functions (set of isomorphism classes of dessins)?
Galois invariants of dessins Question (Grothendieck) When are two dessins in the same Galois orbit? To answer this, it suffices to find a perfect Galois invariant. Three particularly simple Galois invariants are: degree multisets of V 0 and V 1, and the number of edges that bound each face of the dessin equivalently, the cycle types of the monodromy generators very simple to compute! monodromy group of a Belyi function not as simple to compute rational Nielsen class
Precision of known Galois invariants Question How precise is the monodromy cycle type as a Galois invariant? What about other known invariants?
Precision of known Galois invariants not Galois conjugate, but share the same degrees, monodromy groups, and rational Nielsen classes distinguished by a different Galois invariant (due to Zapponi) 3 L. Zapponi, Fleurs, arbres et cellules: un invariant galoisien pour une famille d arbres, Compositio Mathematica 122: 113-133, 2000. 3
The Main Theorem Definition For all positive integers N, let Cl(N) = max n N max λ 1,λ 2,λ 3 n number of Galois orbits of Belyi functions with monodromy of cycle type (λ 1, λ 2, λ 3 ). Our Theorem For all positive integers N, we have Cl(N) 1 16 2 2N 3.
Future Directions Upper bound on Cl(N) Consider Cartesian commutative squares of the form Y X. P 1 P 1 f = (z+1)2 4z For a fixed right morphism, consider all possible left morphisms Leads to an extrinsic Galois invariant for the right morphism (used in the proof of the Main Theorem) How does one define this invariant intrinsically?
Acknowledgements Akhil Mathew, my mentor, for his valuable insight and guidance Prof. Noam Elkies for suggesting this project, and several useful discussions Dr. Kirsten Wickelgren for useful conversations Prof. Curtis McMullen for answering some of my questions The PRIMES Program for making this research possible Thanks to all of you for listening.